Pioneering Physical AI
MGI Tech and Shanghai AI Lab Pioneer Physical AI for Life Sciences
MGI Tech and the Shanghai Artificial Intelligence Laboratory have unveiled two major tools aimed at turning AI into a hands-on lab scientist. The products — ProtoPilot and BioLab Bench — are designed to let AI agents translate research ideas directly into actions on real, automated lab equipment, according to PR Newswire.
The announcement marks what the companies call a new era of 'physical AI for life sciences.' MGI's AI subsidiary, Genoria AI, is leading the push. The research behind both tools was published as a preprint on arXiv, under the identifier arXiv:2606.31763, reports Yahoo Finance.
ProtoPilot Builds a Self-Improving AI Lab Agent
ProtoPilot is a multi-agent system that learns and improves on its own. It is driven by real-world lab scenarios, meaning it gets smarter the more it is used in actual experiments. Think of it as a team of AI workers that plan, adapt, and carry out lab tasks without constant human input, according to PR Newswire.
The system bridges what researchers call the gap between 'digital intelligence' and 'physical execution.' In plain terms, it connects what an AI thinks with what a robot arm or lab machine actually does. That connection has been one of the hardest problems in lab automation to solve.
BioLab Bench Is the First Full AI Evaluation Framework
BioLab Bench is described as the industry's first comprehensive framework for testing AI agents in a lab setting. It scores AI performance across the full chain — from understanding a user's research goal all the way to sending precise commands to lab devices, Yahoo Finance UK reported.
Most AI benchmarks only test one slice of performance, like reading a paper or writing code. BioLab Bench tests the entire pipeline in one go. That makes it a much harder and more realistic test of whether an AI can actually run an experiment from start to finish.
Genoria AI Launches in April 2026 With a Clear Mission
MGI Tech set up Genoria AI in April 2026 as a dedicated AI for Science unit. Its core mission is to build what the company calls 'dry–wet closed-loop infrastructure' for life sciences. In simple terms, that means connecting computer-based AI analysis (dry) with physical lab work (wet) in one continuous loop, according to ADVFN.
The goal is to remove the bottleneck between data and discovery. Scientists currently spend large amounts of time translating digital results into manual lab steps. Genoria AI wants that handoff to be automatic and continuous.
Why Physical AI Could Reshape Biological Research
Physical AI — AI that controls real machines, not just screens — is still rare in biology. Most lab AI tools today help with data analysis or drug design on a computer. Getting AI to reliably run physical experiments is a far harder challenge, and one with massive potential to speed up drug discovery and genetic research.
MGI Tech already makes genomic sequencing equipment used in labs worldwide. Pairing that hardware expertise with Genoria AI's software could give the company a strong edge. PR Newswire noted the launch positions MGI at the frontier of automated biological research.
Sentiment
NeutralPublishers
6
Articles
5
Reach
6
